28.5% of job postings now ask for AI skills. We analyzed 558,596 live job ads.
We scanned the full text of every active posting in the JobLobster corpus — 558,596 live job ads collected directly from employer applicant-tracking systems — to measure how far AI expectations have spread across the job market. The short answer: well beyond tech, and much further up the ladder than down it.
Key findings
- 28.5% of all active job postings mention AI skills (140,101 of 492,144) — across every profession, not just tech.
- AI demand has left the engineering department: 45% of marketing postings, 39% of legal & compliance, 34% of HR, and 33% of sales ads now mention AI — against 44% for software engineering itself.
- Senior roles demand AI three times more often than entry-level roles — 40.3% of senior-IC postings vs 13.1% of entry-level. AI reads as a seniority requirement, not a shortcut for juniors.
- ChatGPT is the most-named tool (10,030 postings), with Claude close behind (8,527), then Copilot (5,439) and Gemini (2,877).
- 41.9% of postings disclose a salary band — and remote employers are ~1.5× more transparent than on-site ones (55.8% vs 38.5%).
- Only 15.9% of listings with a known work mode are fully remote in mid-2026 — the “remote correction” is measurable.
AI skill demand by profession
The stereotype says AI requirements live in engineering job ads. The data says marketing, design, security, and product management are just as saturated — and the professions with the least AI language (healthcare, science, construction) are regulated or physical fields where tooling changes reach job ads last.
| Profession | Share |
|---|---|
| Data & AI | 64.7% |
| Product management | 54.9% |
| Security | 50.7% |
| Design | 50.1% |
| Marketing | 45.1% |
| Software engineering | 43.9% |
| Legal & compliance | 39.3% |
| People & HR | 34% |
| Sales | 32.5% |
| Customer success | 32.4% |
| Finance & accounting | 24.6% |
| Education | 22% |
| Operations | 17.7% |
| Construction & built environment | 11.8% |
| Science & biotech | 8.6% |
| Healthcare | 7.1% |
The seniority twist: AI is a senior-level requirement
The common worry is that AI erodes entry-level work. Job ads tell a sharper story: employers expect senior people to wield AI. A senior individual-contributor posting is three times more likely to mention AI skills than an entry-level one. Judgment plus AI leverage is what the market is pricing, and employers appear unwilling to trust that leverage to people who haven’t built the judgment yet.
| Seniority level | Share |
|---|---|
| Senior individual contributor | 40.3% |
| VP / Head of | 38.1% |
| Director | 33.8% |
| Manager | 29.8% |
| Analyst / IC | 25.8% |
| Entry level | 13.1% |
Which AI tools do employers name?
Most postings ask for AI skills generically, but when employers name tools, ChatGPT leads — and Claude has nearly closed the gap. Mentions of either dwarf every other tool.
| Tool | Postings mentioning it |
|---|---|
| ChatGPT / OpenAI / GPT-4/5 | 10,030 |
| Claude / Anthropic | 8,527 |
| Copilot | 5,439 |
| Gemini | 2,877 |
| Perplexity | 469 |
| Midjourney | 232 |
Salary transparency: remote employers show their numbers
41.9% of active postings publish a salary band. The split by work mode is the interesting part: employers hiring remotely — who compete across state and national lines — disclose pay far more often than on-site employers who compete only with the businesses down the road.
| Work mode | Share |
|---|---|
| Remote | 55.8% |
| Hybrid | 51.1% |
| On-site | 38.5% |
The remote-work map, mid-2026
Fully-remote listings are now 15.9% of postings with a known work mode (on-site: 48.0%, hybrid: 36.1%). Remote work concentrated into a handful of professions rather than disappearing:
| Profession | Share |
|---|---|
| Security | 30.2% |
| Sales | 27.9% |
| Product management | 27.7% |
| Marketing | 27.7% |
| Design | 24.3% |
| Data & AI | 24% |
| Software engineering | 19.5% |
| Finance & accounting | 16.1% |
| Operations | 9.2% |
| Healthcare | 9.1% |
| Construction & built environment | 5.4% |
Median posted salary by profession
Among USD postings that disclose a band, the median midpoint by profession (in thousands):
| Profession | Share |
|---|---|
| Data & AI | 170K USD |
| Product management | 168K USD |
| Security | 165K USD |
| Software engineering | 163K USD |
| Legal & compliance | 163K USD |
| Design | 162K USD |
| Marketing | 136K USD |
| Customer success | 125K USD |
| People & HR | 125K USD |
| Finance & accounting | 118K USD |
| Sales | 116K USD |
| Operations | 109K USD |
| Healthcare | 103K USD |
| Construction & built environment | 101K USD |
| Education | 67K USD |
Methodology
JobLobster maintains a live index of job postings pulled directly from public employer applicant-tracking-system feeds (Greenhouse, Lever, Ashby, Workday and others) — primary-source ads, not scraped aggregator copies. This study covers all 558,596 postings active in the corpus during the first week of July 2026; 492,144 with complete description text were scanned for AI-skill language. A posting counts as mentioning AI skills if its text includes generative AI, artificial intelligence, standalone “AI”, AI-powered tooling, LLMs / large language models, or prompt engineering; classic “machine learning” alone is excluded (it adds just 0.7 points). Professions and seniority come from JobLobster’s deterministic role classifier. Salary figures use only postings that disclose numeric bands. The corpus skews toward North American professional roles; percentages describe this corpus, not the entire global labor market.
“AI skills in job postings, July 2026” — analysis of 558,596 live job postings. Link: https://joblobster.ai/research/ai-skills-in-job-postings-2026. Custom cuts for journalists: hello@joblobster.ai.
Questions
Where does this data come from?
JobLobster continuously indexes live job postings directly from public applicant-tracking-system feeds (Greenhouse, Lever, Ashby, Workday and others). This study scanned the full text of 492,144 active postings — every active posting in the corpus with a complete description — during the first week of July 2026. Nothing here is survey data; every number is counted from what employers actually wrote in their ads.
What counts as an "AI skills mention"?
A posting counts once if its description mentions generative AI, artificial intelligence, "AI" as a standalone term, AI-powered tools, LLMs or large language models, or prompt engineering. Classic "machine learning" alone (an ML-engineering skill, not an AI-tools skill) is counted separately and excluded — it adds only 0.7 points. Tool names are counted by literal mention of the tool or its maker.
Can I cite or republish these numbers?
Yes — cite "JobLobster analysis of 558,596 live job postings, July 2026" and link to this page. For the underlying breakdowns, a chart, or a custom cut of the data (by industry, region, or seniority), email hello@joblobster.ai — we can usually turn custom cuts around within a day.
